# Repo Overview — Where Things Live > **You're a stranger in this repo.** This page tells you which directory holds what, and which file is the right entry point for your role. The branded entry is [`README.md`](README.md); this is the *navigation map*. ## The two frameworks This repo holds two complete, complementary frameworks. They were built independently; they reinforce each other. | Framework | Question it answers | Lives in | | :--- | :--- | :--- | | **Interlace Lint Bench (ILB)** | *Is this lint rule accurate, fast, and stable?* — F1, precision, recall, ms/file, drift across {Node, TS compiler, ESLint, parser, cache} | [`benchmarks/`](benchmarks/) | | **CI/CD Impact** | *What does a CI minute cost the org?* — `$/CI minute`, money × velocity × deliverability axes, falsifiable predictions, vendor-neutral analyzer scorecard | [`cicd-impact/`](cicd-impact/) | ILB measures **rule quality**. CI/CD Impact measures **org-economic value**. Together they let you reason about adoption ROI in finance terms — not just F1 terms. ## Pick your role | You are a… | Start here | | :--- | :--- | | **Developer dropping Interlace into a project** | [`apps/docs/content/docs/launch.mdx`](apps/docs/content/docs/launch.mdx) — the public landing | | **Buyer comparing SAST tools** | [`apps/docs/content/docs/compare.mdx`](apps/docs/content/docs/compare.mdx) — Interlace vs. CodeQL vs. Semgrep vs. Snyk Code | | **AI agent (Claude Code, Cursor) asking what to read** | [`.agent/rules/bench-context.md`](.agent/rules/bench-context.md) — single ~250-line digest | | **Reviewer / researcher** evaluating the bench methodology | [`benchmarks/README.md`](benchmarks/README.md) — 10 principles, vocabulary contract, the 10 benches | | **Finance / engineering manager** asking what the bench is worth | [`cicd-impact/README.md`](cicd-impact/README.md) — `$/CI minute` derivation | | **Contributor opening a PR** | `CONTRIBUTING.md` (planned) — gates, severity contract, vocabulary contract | | **Auditor reviewing project governance** | [`GOVERNANCE.md`](GOVERNANCE.md) — roles, decision rules, conflict policy | | **MITRE / OWASP submitter** | [`benchmarks/audits/INDEX.md`](benchmarks/audits/INDEX.md) — the submission packets | | **Agent / MCP integration** | We retired per-plugin MCP servers. Use the official [`@eslint/mcp`](https://github.com/eslint/mcp) — it auto-discovers every plugin. See [docs/integrations/claude-code](apps/docs/content/docs/integrations/claude-code.mdx). | | **CI engineer** | [`.github/actions/audit/README.md`](.github/actions/audit/README.md) — one-line GitHub Action | ## Doc surfaces — which is canonical for what? | If a claim disagrees, this wins | For… | | :--- | :--- | | [`benchmarks/state.json`](benchmarks/state.json) | Roadmap status (machine-readable) | | [`benchmark-results/scorecard.md`](benchmark-results/scorecard.md) | Live bench numbers (regenerated; never paste numbers into static docs) | | [`benchmarks/lib/result-schema.json`](benchmarks/lib/result-schema.json) | Vocabulary contract (every result file's allowed fields) | | [`benchmarks/README.md`](benchmarks/README.md) | Bench philosophy + 10 principles + 10 benches | | [`cicd-impact/methodology.md`](cicd-impact/methodology.md) | `$/CI minute` formula derivation | | [`scripts/ilb-wild.mjs`](scripts/ilb-wild.ts) (`BENCHMARK_REPOS` array) | Which OSS repos are in the wild corpus | | Per-rule `meta.docs` | The CWE / CVSS / confidence / severity for that rule | ## Three frameworks of "philosophy" Three different doc surfaces use the word *philosophy*; they don't conflict — they sit at different layers: | Doc | Layer | Question | | :--- | :--- | :--- | | [`cicd-impact/value-philosophy.md`](cicd-impact/value-philosophy.md) | **Foundational** — Buffett / Munger / capitalism + humanism | What *is* value? Why is static analysis hard to measure? | | [`cicd-impact/philosophy.md`](cicd-impact/philosophy.md) | **Operational** — three axes (money, velocity, deliverability) | What does CI/CD friction *cost* an organization? | | [`benchmarks/README.md`](benchmarks/README.md) §1 | **Measurement** — 10 principles, FP/FN contracts, vocabulary | How do we make accuracy claims that survive hostile review? | A reader landing on any one is welcome to start there; the other two are linked. ## Strategic surfaces (read for direction) | Doc | Purpose | | :--- | :--- | | [`benchmarks/LEADERSHIP_ROADMAP.md`](benchmarks/LEADERSHIP_ROADMAP.md) | 4-phase plan to lead JS SAST (humans + agents). 67+ items shipped to date | | [`benchmarks/state.json`](benchmarks/state.json) | Machine-readable status of every roadmap item | | [`benchmarks/audits/INDEX.md`](benchmarks/audits/INDEX.md) | Index of dated outreach + readiness packets (MITRE / OWASP / replication / governance / paper) | | [`cicd-impact/predictions-registry.md`](cicd-impact/predictions-registry.md) | Falsifiable predictions; evaluated against later observations | | [`benchmarks/FP_FN_REMEDIATION_TRACKER.md`](benchmarks/FP_FN_REMEDIATION_TRACKER.md) | Active FP/FN triage queue | ## Operational artifacts (read to use) | Surface | What you do with it | | :--- | :--- | | [`packages/-mcp/`](packages/) (×11) | Each ESLint plugin as an MCP server — one per security vertical | | [`packages/eslint-formatter-sarif/`](packages/eslint-formatter-sarif/) | SARIF v2.1.0 output for GHAS / Defender / GitLab | | [`Dockerfile`](Dockerfile) | Multi-arch image at `ghcr.io/ofri-peretz/interlace` | | [`.github/workflows/`](. github/workflows/) | Publish, supply-chain attestation, scorecard, benchmarks, etc. | | [`examples/vulnerable-app/`](examples/vulnerable-app/) | Demo target — one fixture per flagship rule | ## How the directories relate ```text ofri-peretz/eslint ├── README.md branded entry (badges + pitch) ├── OVERVIEW.md this file (navigation map) ├── CONTRIBUTING.md how to land a change ├── GOVERNANCE.md roles + decision rules ├── CODE_OF_CONDUCT.md ├── SECURITY.md ├── apps/docs/ public docs site (launch, compare, integrations) ├── packages/ 16 npm packages: 11 plugin MCPs + plugins + cli + telemetry + ... ├── benchmarks/ ILB framework — measurement │ ├── README.md the bench bible │ ├── BASELINE_MATRIX.md plugin → OSS-repo map │ ├── LEADERSHIP_ROADMAP.md 4-phase plan │ ├── state.json machine-readable status │ ├── audits/ dated outreach + readiness packets (see audits/INDEX.md) │ ├── corpus/ CWE-mapped fixtures │ ├── lib/ shared helpers (toolchain, history, stats, schema) │ └── suites/ per-bench runners (ilb-arena, ilb-cwe-corpus, ilb-wild, …) ├── benchmark-results/ live numbers (regenerated; do not paste into static docs) ├── cicd-impact/ CI/CD Impact framework — economics │ ├── README.md start here for org-economic questions │ ├── value-philosophy.md foundational │ ├── philosophy.md operational (three axes) │ ├── methodology.md $/CI minute formula derivation │ ├── analyzer-evaluation-framework.md vendor-neutral scorecard │ ├── predictions-registry.md falsifiable predictions │ ├── data/ real Actions-API data │ ├── outputs/ computed outputs (unit-cost.json, …) │ └── scripts/ fetch + compute + render scripts ├── examples/vulnerable-app/ demo target (one fixture per flagship rule) ├── scripts/ ilb-* + audit-* + check-* + sync-* + ... └── .agent/ agent-orchestration (rules, workflows, skills, agents) └── rules/bench-context.md ← single-read prelude for AI agents ``` ## What this doc is NOT - Not the pitch — that's [`apps/docs/content/docs/launch.mdx`](apps/docs/content/docs/launch.mdx). - Not the bench bible — that's [`benchmarks/README.md`](benchmarks/README.md). - Not the CI-cost philosophy — that's [`cicd-impact/`](cicd-impact/). - Not the changelog — that's automated via Changesets per package. If a claim in this doc disagrees with a doc it points to, the linked doc wins. This page only routes; it doesn't author.